From: Brett Nordgren 
To: psnlist 
Sent: Thu, 10 Oct 2013 16:03
Subject: Re: OpAmp noise


Chris,
>You wrote:
>****The two chip chopper amp design gives such low noise that
>in any circuit with a source resistance greater than about 500
>Ohms, the intrinsic input resistor noise will dominate the overall
>circuit noise. And Chopper amplifiers DON'T show ANY 1/f noise !
>THIS circuit DOES use +/-15 V SUPPLIES !

    But if we are monitoring a sensor having
significantly higher resistance than 500 Ohms which, itself, may have
some modest amount of 1/F noise, it's not clear what that added
complexity buys us.  When Dave tried replacing the AD706 with an
LTC1151, which also should have zero 1/f noise, the instrument
performance did not visibly improve.
****The LTC1151 has ~no 1/f noise, BUT it still has quite a LOT of
ordinary noise and it is about comparable with the AD706, maybe a
bit worse. I would not expect there to be much advantage in using it.
    What were comparable FFTs of the two configurations, please ?
You may not see much from a visual inspection of the traces.

>****I am puzzled as to what you are quoting when you say your
>position sensor has an OUTPUT resistance of 50 K ? I thought that
>you / we had the switched capacitors connected to an opamp ? Is this
>impedance resistive or purely reactive ?

There was a time when I, too, believed that switched capacitor
circuits were inherently noiseless, the only noise coming from such
things as imperfect switches.  However I have since found that's not
true.  A simple switched capacitor circuit exhibits an apparent
resistance =3D 1/(f C), which resistance will exhibit the noise
associated with a discrete resistor of that value.
****See http://bnordgren.org/seismo/FBV/FBV_Circuit_Description.pdf
Sorry, but I still don't understand. Where were you measuring this=20
noise
please ? Across C4 on T8 of the LTC1043, or on the T1 output of the
AD706, U1A ? The bridge runs at 62.5 K Hz and has a capacitance of
48.2 pF, giving a REACTANCE of 52.8 K  - this is NOT a resistance !
As far as I know, NEITHER Capacitors NOR Inductors generate any
significant noise - both devices may have a small inherent resistance.

The 52k value I used is actually the sensor DC output resistance,
both as measured and simulated in Spice.  Whether that has a
corresponding sqt(4kTR) voltage noise density, I don't know yet, but
to be conservative I was assuming that it did.  In any case, it
definitely does affect the degree to which the OpAmp's current noise
adds to its total low-frequency noise.  I'll try some more
simulations to see if I can discover how that 52k varies with=20
frequency.
****???? Where does ANY opamp's current or other noise fit in ? Are
you leaving UA1 in circuit while you are measuring the noise ? Might it
be a good idea to make up a William's ultra low noise type amplifier
just to test the circuit noise ?
    I thought that the noise was sqrt(4kTRB) where B is the bandwidth ?

Hi Brett,

     Sorry to appear a bit thick, but I don't know exactly what
measurements you are taking, how you are doing them, or with what
equipment. With several noise sources, you add the squares of the
noise voltages and take the square root the sum.

    Regards,

    Chris Chapman



__________________________________________________________

Public Seismic Network Mailing List (PSNLIST)

To leave this list email PSNLIST-REQUEST@.............. with 
the body of the message (first line only): unsubscribe
See http://www.seismicnet.com/maillist.html for more information.